Fire Investigation in Italian WasteTreatment Plant: lessons learned - - PowerPoint PPT Presentation

www.trr.it or download our APP on

Fire Investigation in Italian WasteTreatment Plant: lessons learned and future development

• Ing. Alfredo Romano (General Manager TRR S.r.l.)
• Ing. Giovanni Romano (General Manager RSM S.t.p.)

TRR is a private, independent consultant company that operates in Health, Safety and Environmental field since 1984. Its headquarter is in Italy. ✓ Hazard and Operability Analysis ✓ Audit & Training (Existing Plant) ✓ Quantitative Risk Assessment (QRA) ✓ Safety and Integrity Analysis ✓ Fire prevention (Fire Engineering) ✓ Health, Safety and Environmental Engineering ✓ Modelling and software

Who we are Where we operate

✓ Oil & Gas, Gas and Oil Field ✓ Storage and Refining ✓ Iron & Steel Industry ✓ Chemistry ✓ Transport of dangerous goods

Our service TRR (TECNOLOGIA RICERCA RISCHI)

SCOPE

❑In June 2017, in the north of Italy, a serious fire occurred in a waste treatment plant (entire warehouse was involved) ❑The scope of this intervention is to present fire modeling in order to establish if the sequence of events leading up to the accident could be suitable with developing times of the fire event ❑Results from the simulation have been compare with information and evidences collected on field

• 1. Fire Event description
• 2. Fire Data collection
• 3. Fire Modeling
• 4. Conclusions about incident investigation
• 5. Future Development

SUMMARY OF THE PRESENTATION

• 1. EVENT DESCRIPTION

❑Warehouse site near industrial area (petrolchemical plant) ❑Warehouse surface: 2960 m2 ❑Warehouse height: 12 m ❑ Fire duration: 5 hours

INTERNAL EMERGENCY TEAM t= 180 s 6 6 1

3 5 4

2 7 Matress Other material (various)

triturated

Tyre Box triturated triturated

triturated

wood wood FIRE6 (t = 180s)

IGNITION t = 0 s

TRITURATED MACHINE FIRE6 (t = 300s)

MACHINE

FIRE 1 (t=600s) FIRE 3 (t > 600s) FIRE 5 (t>600s) FIRE 4 (t > 600s) FIRE2 (t>600s) FIRE (t = 300s) FIRE 7 (t>600s)

FIRE BRIGADE EMERGENCY TEAM t= 600 s

• 2. FIRE DATA COLLECTION

❑Personell interview (after fire event) ❑Fire Brigade and Local Authorities official report ❑Management information and interview

FLASH OVER (t > 900 s)

QUESTION ABOUT FIRE EVENT a) the fire in the warehouse was congruent in terms of propagation speed and intensity? b) if the fire in the outside container of combustible material was suitable with the development of the fire inside the warehouse ? c) probable causes that led to the fire?

• 3. FIRE MODELING (Questions about)

6 6 1 3 5 4 2 7

Wood Wood

triturated

triturated Matress Miscellaneous triturated Tyre Box triturated

• 3. FIRE MODELING (layout&material parameters)

Conveyor Belt (out of service)

• 3. FIRE MODELING (3d layout)

6 m 10 m 100 m 25 m

Material Density (kg/m3) Auto-ignition temperature (°C) HRRPUA1 (kW/m2) Tritured 50 250 500 Wood 640 450 200 Tritured Wood 640 400 200 Miscellaneous 900 350 300 Rubber 1400 400 500 (1) HRRPUA: Heat Release Rate Per Unit Area

• 3. FIRE MODELING (material parameters)

Material Density (kg/m3) Heat conductivity (W/m K) New Jersey barriers 2280 1,8 Sandwitch panels 40 0,02 Polycarbonate glasses 1380 0,2

❑ Software used: FDS (NIST) /Pyrosim (Thunderhead Engineering) ❑ 400.000 Mesh element ❑ Parallel CPU usage ❑ Machine type: Intelcore i7- 8 core ❑ Simulation time (machine): 90 hours

• 3. FIRE MODELING (software&tools)

FDS employs the material characteristics of the elements of the survey domain, in order to estimate the growth and spread of the fire The ability of the FDS model to predict accurately the temperature and speed of the gases developed by the fire was previously evaluated by the NIST by conducting appropriate experiments, both on a reduced scale and in a real scale and the subsequent analysis of the data from the field.

• 3. FIRE MODELING (results)

a) the fire in the warehouse was congruent in terms of propagation speed and intensity?

• 4. CONCLUSION

YES

147 s

Time 200 s: the fire was extended to the compartment 6/6 near the pile of combustible material

• 4. CONCLUSION

Time 300 s: the fire was extended to the compartment 1 (matress)

• 4. CONCLUSION

322 s

• 4. CONCLUSION

Time >1000 s: the fire was extinguished – the warehouse not collapsed

1002 s

b) the fire in the outside box (tyre/rubber) was compatible with the development of the fire inside the warehouse?

• 4. CONCLUSION

NO

• 4. CONCLUSION

❑The development of the fire is congruent with the available elements. ❑The temperature profile detected with a virtual plane on the tire caisson leads to low temperature values not able to develop the degradation of the rubber, in

• rder to facilitate ignition.

❑The fire continues even if the piles of crushed material to the next compartments begin to decrease in consistency and the flames are lowered with consequent reduction of the temperatures

6 6 1 3 5 4 2 7

Wood Wood triturate d triturated Matress Materiale da cernita triturated triturated

• 4. CONCLUSION

10 m

Conveyor Belt (out of service) Tyre box

c) causes that led to the fire?

• 4. CONCLUSION

UNDER INVESTIGATION

Potential ignition source: flame rocket in tritured material (shredder waste) ? 4. CONCLUSION

• 4. CONCLUSION

❑ Potential ignition source: combustibile and flammable material inside tritured material ❑ Further investigation to be completed

❑ In the last few years, many similar fires occurred (about 260 fire event) in Italy and, for that reason, the Italian Ministry of Environment has recently issued a first version of a “Guideline” according to risk management. ❑ These accidents and the results of the analyses will probably lead to a review of fire-prevention design criteria of these plants. ❑ During the analysis it was found that, at a design level, some measures could be taken, such as for example passive protection or the insertion of a dedicated sprinkler system, even limited to a portion of the warehouse. ❑ The fire and/or smoke detection could be considered even if the intervention of the team was timely and the high fire risk processes are carried out during the personnel supervision. Regarding the modeling of the fire, the experimental starting data of the materials should be analyzed, in order to better characterize the phenomenon. 5. FURTHER DEVELOPMENT

LASER SCAN 3D POINT CLOUD 3D MODEL IN POINT CLOUD (bubble view) FIRE SIMULATION SOFTWARE (PYROSYM/FDS)

5. FUTURE DEVELOPMENT

www.trr.it or download our APP on

Thanks for your attention

And Special Thanks to

• Ing. Alfredo Romano
• Ing. Giovanni Romano